AI/diffusers
1
0
Fork 0
mirror of https://github.com/huggingface/diffusers.git synced 2026-01-27 17:22:53 +03:00
Code Activity

[SVD] support generators that are created on GPU (#6484)

Browse Source 
* debug generator

* fix?

* fix?

* fix

* remove print.

* revert none check
This commit is contained in:
Sayak Paul
2024-01-11 20:08:18 +05:30
committed by GitHub
parent 9df566e6da
commit 5dc3471380
1 changed files with 8 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
src/diffusers/pipelines/stable_video_diffusion/pipeline_stable_video_diffusion.py
View File

@@ -429,15 +429,20 @@ class StableVideoDiffusionPipeline(DiffusionPipeline):
fps = fps - 1
# 4. Encode input image using VAE
image = self.image_processor.preprocess(image, height=height, width=width)
noise = randn_tensor(image.shape, generator=generator, device=image.device, dtype=image.dtype)
image = self.image_processor.preprocess(image, height=height, width=width).to(device)
noise = randn_tensor(image.shape, generator=generator, device=device, dtype=image.dtype)
image = image + noise_aug_strength * noise
needs_upcasting = self.vae.dtype == torch.float16 and self.vae.config.force_upcast
if needs_upcasting:
self.vae.to(dtype=torch.float32)
image_latents = self._encode_vae_image(image, device, num_videos_per_prompt, self.do_classifier_free_guidance)
image_latents = self._encode_vae_image(
image,
device=device,
num_videos_per_prompt=num_videos_per_prompt,
do_classifier_free_guidance=self.do_classifier_free_guidance,
)
image_latents = image_latents.to(image_embeddings.dtype)
# cast back to fp16 if needed